CONFERENCE SUMMARY
Recent Developments in
Credit Scoring:
A Summary
October 24, 2006
www.philadelphiafed.org
Recent Developments in Credit Scoring Recent Developments in
Credit Scoring:
A Summary
Marvin M. Smith, Ph.D.
Federal Reserve Bank of Philadelphia
Acknowledgment: I would especially like to thank Christy Chung Hevener. This symposium
would not have been possible without her untiring efforts.
The views expressed here are those of the author and do not necessarily represent the views of the Federal
Reserve Bank of Philadelphia or the Federal Reserve System.
www.philadelphiafed.org
Recent Developments in Credit Scoring TABLE OF
CONTENTS
I. Introduction.......................................................................................................5
II. Overview of Credit Scoring................................................................................5
Credit Reporting..........................................................................................................5
Credit Score........................................................................................................6
Discussion.............................................................................................................8
III. FICO Expansion Score.......................................................................................8.
Discussion...........................................................................................................11
IV. VantageScore...................................................................................................12.
Discussion...........................................................................................................14
V. HealthScore.....................................................................................................14.
Discussion...........................................................................................................16
VI. Concluding Note..............................................................................................17.
Exhibit 1: Symposium Agenda...................................................................................19
Exhibit 2: Institutions Represented at the Symposium.............................................21
1
www.philadelphiafed.org
The synopsis of the keynote address was written by Payment Cards Center Industry Specialist James C. McGrath.
Recent Developments in Credit Scoring The symposium drew 140 people representing local banks, nonprofit organizations, and local government
agencies. The featured guest speakers were Patricia Hasson from Consumer Credit Counseling Service
of Delaware Valley, Janice Horan from Fair Isaac Corporation, Maxine Sweet from Experian, and Jeremy
Blackburn from Canopy Financial.
Recent Developments in Credit Scoring
www.philadelphiafed.org
I. Introduction
Credit scores are important indicators of
consumers’ credit profiles, and they are used by
mortgage lenders, credit card issuers, and many other
financial institutions to assess consumers’ willingness
and ability to repay their financial obligations. Over
the past decade, they have become the main factor
that determines borrowers’ access to capital and the
cost that borrowers will ultimately bear. But how do
consumers make sense of three credit scores, one
from each of the three major credit reporting agencies? What about those consumers who have little or
no credit history and thus no traditional credit score?
And what about consumers with high-deductible
health plans who need credit to pay for costly medical procedures?
To shed some light on these questions, the
Community Affairs Department of the Federal Reserve Bank of Philadelphia hosted a symposium entitled “Recent Developments in Credit Scoring” on
October 24, 2006. This symposium was conceived
in order to explore developments in the industry
that are making credit scoring a more accurate and
predictive tool. The symposium highlighted three important developments that address the concerns outlined above by refining or augmenting current credit
scoring techniques.
II. Overview of Credit Scoring
Patricia Hasson of the Consumer Credit
Counseling Service of Delaware Valley opened the
symposium by providing an overview of credit scoring. Hasson traced its origins and offered a synopsis
of the current credit scoring system, which, she
stressed, is designed to make lending and borrowing
more fair and efficient. She also alluded to the industry experts who would follow her on the program to
discuss the latest innovations in credit scoring — innovations devised to further improve the process for
lenders while further leveling the playing field for
borrowers.
www.philadelphiafed.org
Hasson began, however, by drawing on her
experience as a commercial lender earlier in her career. She recalled originating loans that ranged from
automobile to home equity to unsecured consumer
loans. What was especially memorable to her was
looking in one of her first files with a credit history
dating back to the 1960s and finding interviews with
the applicant’s neighbors, attesting to his character.
She mused that not many loans would be made today
if lenders based their decisions on the character references of someone’s neighbors. Clearly, such subjective
information, according to Hasson, is an ill-advised
way to gauge someone’s creditworthiness today.
Credit Reporting
Turning to the origins of credit reporting
and the credit scoring system, Hasson pointed out
that credit reporting began just over 100 years ago. It
started with small merchants sharing information on
their customers’ payment patterns and the likelihood
that a customer would repay the loan. This was an
expedient way of mitigating risks. These collaborative efforts led to the establishment of merchant associations, which, with the advent of computerization
and through consolidation, developed into larger
credit bureaus. However, the system was far from
advantageous to borrowers whose information was
maintained by these bureaus. The bureaus lacked
safeguards for ensuring the privacy of the information
they held on consumers. They also recorded some
types of information that were not good predictors of
a consumer’s creditworthiness. Moreover, consumers
were unable to access the information credit bureaus
had collected about their credit behavior.
Hasson noted that these shortcomings
weren’t addressed until Congress passed the Fair
Credit Reporting Act (FCRA) in 1971. The FCRA
delineates standards governing consumer privacy and
accuracy in reporting. As a result of the act, credit
reports began including positive financial information. Furthermore, consumers were afforded the right
to obtain copies of their credit reports and dispute
information they believe is inaccurate.
Recent Developments in Credit Scoring Credit Score
According to Hasson, most people associate
credit scores with the FICO score. In fact, the Fair
Isaac Corporation, the developer of the FICO score,
began its path-breaking work with credit scoring in
the late 1950s. Fair Isaac built its first credit scoring
system for American Investments in 1958. In the
early 1960s, the company built a credit scoring system
for Montgomery Ward and later developed a credit
scoring system for a bank credit card for Connecticut
Bank and Trust (1970). These early scoring systems
were specific to each company. Hasson added that
the first general-purpose FICO score was developed
in 1989 in partnership with the credit bureau Equifax
and continues to be a popularly used score today. She
observed that the FICO model has become the standard most widely used for determining consumers’
credit risk.
Fair Isaac developed a mathematical algorithm that uses financial information on a consumer
to calculate a credit score. This score, in turn, is
compared with the scores of other consumers with
similar financial profiles, along with their repayment
behavior, to determine a consumer’s creditworthiness.
Hasson pointed out that the use of a credit score
makes it possible to receive an instant response for
a credit request at retail stores or to get a 30-minute
approval on home equity lines. However, she cautioned that while a credit score is an assessment of
risk, it is only one of many factors considered when
evaluating whether someone is granted credit; other
factors might include income or collateral. But a
credit score is increasingly being used to determine
the terms of consumer credit, such as the interest rate
on a loan.
Hasson briefly discussed the basic components of a credit score. She indicated that a credit
score is based on the information contained on a
consumer’s credit report, which is broken down into
five categories. Each category contributes a certain
percentage to the total score. Generally speaking,
payment history accounts for 35 percent, outstanding
debt for 30 percent,1 length of credit history for 15
percent, new credit 10 percent, and how many and
what types of credit used (installment versus revolv-
Recent Developments in Credit Scoring
ing credit) another 10 percent. The more favorable
the information in each category, the more points
that category contributes to the total score. Scores
range from 300 to 850, and consumers with higher
scores receive better credit terms. Hasson noted that
the breakdown of credit scores across the nation
reveals that the largest percentage of Americans (29
percent) fall into the 750 to 799 range.
Hasson then identified ways that might help
improve a credit score, including paying down credit
card and other debt, making payments on time, refraining from closing credit card accounts that have
been held for a long period of time and that help to
establish a long payment history, avoiding opening
new accounts, and, in general, establishing a longterm, positive credit history. She also warned of actions that would hurt a credit score, including missing payments, using a credit card to its maximum limit, excessive shopping for credit, and having a greater
number of revolving loans compared to installment
loans. But she hastened to add that adopting a strategy to improve a credit score is an individual matter
and there isn’t a “one size fits all” solution. Hasson
thought that it would be prudent to consult someone
with experience with and knowledge of credit scoring
for assistance in charting a course of action.
Hasson emphasized that the credit scoring
system has been beneficial to both creditors and borrowers by creating efficiency in the marketplace. She
recalled a time when only the wealthy were able to
obtain loans. Now access to credit by borrowers of
different income levels has increased. The automation of financial information and credit decisionmaking has made possible a faster extension of credit
at cheaper prices. Furthermore, the Internet affords
borrowers the opportunity to look beyond their
neighborhood and seek the best credit terms nationwide. Hasson also suggested that the scoring system
has leveled the playing field for those who have ex-
This category has an inverse relationship between the amount
of debt and the number of points contributed to the credit score —
namely, the lower the debt, the higher the points and vice versa.
1
www.philadelphiafed.org
perienced financial difficulties in the past. It enables
them to repair their credit and raise their score in
order to seek better credit terms.
However, Hasson pointed out that lenders
are not the only ones using credit scores today.
Others, such as landlords, utility companies, and potential employers, also rely on them. Hasson referred
to a 2005 report by Spherion, a U.S. recruiting firm,
which revealed that the number of employers using
credit scores and credit reports to determine the suitability of a candidate had increased 55 percent over
the previous five years.2
Given the important role that financial behavior and the three little numbers in a credit score
play in our lives, Hasson called attention to the need
for consumers to be more aware of their financial profile. She pointed to a recent survey by Capital One,
which disclosed that 27 percent of the consumers surveyed had never looked at their credit report.3 Here,
she noted that consumers have been aided by the Fair
and Accurate Credit Transactions (FACT) Act of
2003. In addition to providing more safeguards regarding credit reporting and provisions governing identity
theft, the FACT Act also granted all Americans the
right to obtain a free copy of their credit report annually from each of the three national credit reporting
agencies (Equifax, Experian, and TransUnion). But
consumers must pay a fee to obtain their credit score.
However, Hasson wondered whether consumers who
do obtain a copy of their credit report are able to understand it and use it to improve their financial situation. On this point she touted the virtue of financial
education and making use of the services offered by
credit or housing counseling organizations. Hasson
also challenged consumers to be more proactive when
it comes to their credit report and credit score. She
emphasized that a credit score is only as good as the
data it is based on. Errors on a credit report (which
are reflected in the credit score) can adversely affect a consumer’s ability to obtain credit or favorable
credit terms. Hasson urged consumers to be vigilant
in monitoring their credit report for errors and taking
corrective action when errors are detected.
Hasson’s view of the impact of the credit
scoring system on the financial landscape is that it is
both a blessing and a curse. While the scoring system
has greatly eased the process of obtaining credit, that
very ease of access to credit has contributed to our
becoming a nation of debtors, not savers. She observed that credit card debt in the U.S. exceeds $820
billion, while total nonmortgage debt is roughly $2.2
trillion.4 She also cited a study by the U.S. General
Accountability Office that reported that late fees
on credit cards increased from $13 in 1995 to $34
in 2005, while interest rates charged for late payments and exceeding credit limits increased more
than 30 percent.5 Furthermore, 55 percent of college students obtain a credit card in their freshman
year, and the average college graduate leaves school
with $2,700 in credit card debt, not including student loans. Hasson thought that promoting financial
literacy earlier in people’s lives might address the
overall debt situation and help decrease the number
of delinquencies, bankruptcies, and other financial
difficulties that adults may encounter.
All told, Hasson thought that improvements
are being made in the credit scoring system. She alluded to one of the innovations — discussed by a
later speaker — that gives consumers who historically have not been included in the traditional credit
system an opportunity to obtain credit scores. She
held out hope that other innovations might also allow businesses to expand and extend their financial
services with lesser risk involved.
These figures have since changed. As of March 2007, they
are $888.2 billion and $2.4 trillion, respectively. See http://www.
federalreserve.gov/releases/g19/current/default.htm.
5
U.S. General Accountability Office, Credit Cards: Increased
Complexity in Rates and Fees Heightens Need for More Effective
Disclosures to Consumers, GAO-06-929.
4
Findings from Spherion Emerging Workforce Study, Spherion
Corporation, 2006.
3
See the Second Annual America’s Financial IQ survey,
October 5, 2006, at http://www.consumer-action.org/press/articles/
second_annual_americas_financial_iq_survey/.
2
www.philadelphiafed.org
Recent Developments in Credit Scoring Discussion
In the discussion that followed Hasson’s
presentation, symposium participants focused on
several aspects of the credit report and its companion
credit score. A number of attendees expressed concern about the price of obtaining a credit score and
that the fee may be prohibitive for a segment of the
population. Although the FACT Act of 2003 allows
consumers to obtain one free credit report each year,
obtaining the credit score that corresponds with that
report can cost about $14.95 for an individual score
or $44.85 for the three scores from the three major
credit reporting agencies, according to a representative from Fair Isaac. While the industry currently
has no plans to offer free credit scores, Hasson suggested that consumer groups could place the issue
on their agendas.6 Also, Fair Isaac is willing to work
with groups that require discounts in cases where
the costs of obtaining credit scores are prohibitive.
However, charging a reasonable fee for a credit score
is permissible under the FACT Act because Congress
acknowledges that credit bureaus incur costs in obtaining data and producing a score.
Audience members were also concerned
about the number and types of inquiries on an
individual’s credit report and how they may affect an
individual’s score. As with most issues concerning
credit scoring, the impact of inquiries depends on the
individual’s credit file. There might be little or negligible impact for those with a long and established
credit history but a more substantial impact for those
with a relatively short or new credit history. In general, inquiries that demonstrate that the individual
is seeking credit are regarded as “hard” inquiries and
will have a negative effect on a credit score.7 How-
In fact, one of the presenters mentioned that support for
that position was echoed by Liz Pulliam Weston, a nationally
syndicated columnist who has written on the topic and
suggested that Congress should provide free credit scores once
a year.
See http://articles.moneycentral.msn.com/Banking/
YourCreditRating/DemandYourFICOScoreNow.aspx
7
Although inquiries remain on a credit report for up to two
years, many creditors generally ignore those inquiries that have
been on the report for six months or more.
6
10 Recent Developments in Credit Scoring
ever, because the FICO model acknowledges the
importance of comparison shopping, inquiries made
within a two-week period are considered a single
inquiry. Also, when the FICO model is constructing
the score, it doesn’t consider inquiries made in the
last 30 days. Unlike hard inquiries, “soft” inquiries
do not get factored into an individual’s credit score.
Examples of soft inquiries include a consumer’s request for his or her credit report, lenders’ requests for
“pre-approved” credit offers, and requests from prospective employers.8
Audience members also raised concerns
about removing inaccuracies from a credit report
and the effect of these actions on the credit score.
In particular, audience members said that they had
not immediately seen the expected increase in their
credit score after negative information was corrected
and, in some cases, had even seen a decrease in their
credit score. The speakers stressed the importance of
allowing enough time for the changes in the credit
report to flow through the system, which can take up
to a minimum of a month for the necessary changes
to be updated and shared with the relevant parties. If
inaccuracies persist, the major credit bureaus should
be contacted, since they are the repositories of credit
data. Hasson also cautioned against comparing scores
from different credit bureaus, since these scores may
differ because of differences in the underlying data
and credit scoring methods.
III. FICO Expansion Score
Janice Horan of the Fair Isaac Corporation
made a presentation on the FICO Expansion Score.
She explained that the FICO Expansion Score was
created to serve the segment of the population who
were unable to obtain a credit score calculated from
traditional credit data. Members of this credit-underserved market fall into two groups. One group consists of those individuals about whom credit bureaus
Soft inquiries also include requests by current lenders for a
review of a consumer’s account and requests by landlords for
purposes of screening tenants.
8
www.philadelphiafed.org
have no information and hence no record; they are
dubbed “no-hit” individuals. The other group is composed of those individuals about whom credit bureaus
have very little information; they are called “thin-file”
individuals. Horan noted, however, that thin-file individuals may have extensive payment histories, but
their thin-file status may be the result of making payments to creditors who do not report payment history
to the traditional credit reporting agencies.
Fair Isaac estimates that about 50 million individuals in the United States, or one in four adults,
are in the underserved credit market. The company
arrived at this number by subtracting the number of
individuals with a full credit record (160 million) from
the adult population in the U.S. as of the last census
(210 million). Of the 50 million in the underserved
credit market, Fair Isaac estimates that 30 million are
thin-file and 20 million are no-hit individuals.
Horan further described the underserved
population as a nonhomogeneous group, consisting of
immigrants, minorities, students, recently divorced or
widowed individuals, and that segment of the population, such as the elderly, that is culturally resistant to
credit and which continues to embrace cash.
According to Horan, the underserved credit
market presents both a problem and an opportunity.
On the one hand, individuals with little or no credit
history at the traditional credit bureaus (and hence
less likely to have FICO scores) are typically left out
of the mainstream lending programs, are charged high
rates if they can get credit products at all, and end
up as part of the underserved credit market. On the
other hand, lenders look upon this segment as an opportunity to generate significant revenue through the
marketing of their financial products while providing
much deserved credit to an underserved population. In fact, Horan indicated that lenders took the
initiative and approached Fair Isaac for assistance in
tapping that portion of the underserved market for
potential low-risk customers.
In response to requests by lenders, Fair Isaac
developed the FICO Expansion Score, which is a
credit risk score based on alternative credit data. The
challenge was to identify the types of data available
www.philadelphiafed.org
on the experiences of the underserved population in
nontraditional credit-granting institutions, scale the
data in a manner similar to that used in the classic
FICO score, develop a score comparable to the classic FICO score, and offer the same score range as
the classic FICO score. Moreover, the FICO Expansion Score had to have the same consumer-friendly
features that the classic FICO score has, namely,
sufficient information to indicate which sources
were used and which data were gathered to assess an
individual’s payment history, statements that explain
why an individual’s score is not higher, a built-in and
operational consumer dispute process, and a scoring method that fully complies with regulations (i.e.,
meets the requirements of the Fair Credit Reporting
Act, Gramm-Leach-Bliley Act, Equal Credit Opportunity Act, and Fair and Accurate Credit Transactions Act).
Horan further explained that the FICO Expansion Score is derived from data that reflect the
consumer’s bill-paying performance on obligations
with credit characteristics but such performance
is not currently reported to the traditional credit
reporting agencies. The desired data sources are national in scale and contain large volumes of records
that offer both positive and negative information.
These data are supplemented with other information,
if available, including thin-file information from the
credit reporting agencies, information from public
records, and application data with characteristics
that may have predictive value.
According to Horan, Fair Isaac recognized
that a new credit score introduced in the market
would have to be able to demonstrate that it can
perform on par with proven scores, such as the classic FICO score. In anticipation of such a challenge,
Fair Isaac, she reported, subjected the FICO Expansion Score to many tests, including retro-validation
and key performance measures in order to ensure its
viability as a predictive product in the marketplace.
Among the main performance metrics lenders are
concerned about are the product’s scorability — the
percent of the lender’s target population for which
the product will provide scores; score distribution
— the score’s ability to sort the population into varying levels of risk (captured in different score ranges),
Recent Developments in Credit Scoring 11
which will allow the lender to distinguish low-risk
consumers from high-risk ones; score rank ordering
— a score that reflects the fact that people with higher scores should perform better than those with lower
scores; and score alignment — the ability of any FICO
Expansion Score (say, 700) to reflect the same risk
assessment as its identical (700) classic FICO score.
One approach to gauging all of these factors
is to undertake a retrospective validation (or retrovalidation). Retro-validation uses accounts that have
been on the books for a period of time because the
accounts’ performance is already documented. The
FICO Expansion Score is used to score the account
holders, and then the score’s predictive ability is
observed by scrutinizing the actual payment performance of the accounts. The score’s efficacy can be
determined by charting its performance with respect
to the performance metrics mentioned above. Essentially, you are posing the following question: If you
had had the FICO Expansion Score at the time these
accounts were opened, would it have accurately predicted their performance? Horan reported that this
procedure was undertaken with the cooperation of
several companies from the credit card, auto finance,
and mortgage industries, including American Express
and HSBC from bankcard issuers; Daimler-Chrysler, Drive Time, and Ford Motor Credit from auto
finance lenders; and Credit Suisse, First Franklin,
Freddie Mac, HSBC Mortgage, and Option One from
the mortgage industry.9 Thus, Fair Isaac used application information from January 2004 to June 2004
to score account holders using the FICO Expansion
Score and then observed the accounts’ payment performance through the spring of 2006, approximately
24 months. The performance criterion for the FICO
Expansion Score was the likelihood that an account
would be delinquent 90 days or more within a 24month period.
In the scorability analysis, Horan indicated
that two criteria were evaluated: whether a consumer
would be scorable and whether the FICO Expansion data sources supplied sufficient information to
generate a predictive score. She reported that the
scorability rates varied by, but were not limited to,
the industry examined, the product considered, and
the marketing channel that served as the source of
the original application. The overall scorable rate for
consumer account holders across all of the participating companies was 68 percent. The auto companies registered the highest scorable rate, while the
mortgage companies had the lowest. According to
Horan, this was primarily because Fair Isaac essentially had access to all of the application data from
the auto companies, but the compilation of rental
histories is highly segmented and all of the data are
not included in the national repositories Fair Isaac is
able to access.10
Horan thought that the resulting score
distribution of the thin-file and no-hit underserved
population over the credit score ranges was quite
revealing.11 The credit scores were distributed fully
across the risk spectrum. This means that there were
applicants in the underserved market who were a
low credit risk and who had the potential to qualify
for prime credit products with lower prices and fees
than are generally available to those in this underserved market. Horan noted that this demonstrated
the possibility that lenders in the three participating
industries (credit card, auto finance, and mortgage)
could increase their portfolios by marketing their
prime products to the low-risk consumers in the underserved market. Moreover, the FICO Expansion
Score would also enable lenders to price the riskiness
of other consumers in the underserved segment.
Horan stressed that part of the philosophy behind
the FICO Expansion Score is that once a consumer
receives a mainstream credit product, the lender is
expected to start reporting the consumer’s credit
She further pointed out that many of the mortgage programs
— particularly subprime and nontraditional — already manually
review rental histories.
11
The credit score ranges are: not scorable; 300-549; 550-599;
600-619; 620-623; 640-649; 650-659; 660-679; 680-699; 700749; and 750-850.
10
Horan indicated that other lenders also participated and their
outcomes were reflected in the results she reported.
9
12 Recent Developments in Credit Scoring
www.philadelphiafed.org
behavior to a mainstream credit reporting agency.
Thus, the FICO Expansion Score would not create
a two-tiered system but would help consumers make
the transition into the mainstream credit market.
Consequently, the next time that consumer applies
for credit he or she can receive a classic FICO score
from Equifax, Experian, or TransUnion.
The final stage of the validation process is to
consider the rank ordering of risk across the range of
credit scores. In the case of the credit card industry,
for a FICO Expansion Score in the 530 to 559 range,
the odds of favorable repayment were 2.6 to 1 — or
for every 3.6 applicants considered, 2.6 of them are
going to pay as previously agreed and one is not.
Similarly in the 700 to 749 range, the odds were 22
to 1. Thus, the rank ordering of risk had the desirable
features of having a significant difference in the odds
between the highest and lowest score bands, and the
progression of odds increased as the score increased.
The final test was to compare the rank ordering ability of the FICO Expansion Score with that of the
classic FICO score. Fair Isaac found that the rank
ordering of the FICO Expansion Score aligned quite
well with the ranking in the classic FICO score.
Horan closed by pointing out that the FICO
Expansion Score complies with all regulatory requirements and provides consumers with the appropriate
disclosures comparable to those that accompany the
classic FICO score, as well as a copy of their Expansion credit report.12
Discussion
The discussion of the FICO Expansion
Score focused on clarification of its use and availability. One symposium participant asked which
lenders were currently using the FICO Expansion
Score. Horan responded that the FICO Expansion
Score has been in existence for approximately three
years and its greatest use has been in the credit card
This also includes reason codes (and explanations of the
codes), which are required by law and indicate where consumers
lost points in their overall Expansion score.
industry, followed closely by the automobile industry. She expressed hope that the FICO Expansion
Score will gain greater acceptance in the mortgage
industry, especially once the secondary market and
the government sponsored enterprises (GSEs) give
mortgage lenders guidance regarding acceptance of
this risk score by those entities.
Another attendee wanted to know when a
lender should use the FICO Expansion Score rather
than the classic FICO score. Horan indicated that
such a decision would typically be made by a credit
policy officer in a company’s home office rather
than at the underwriter level. Presumably, companies would establish a policy that outlines the criteria that must be met in using one of the two scores.
Horan speculated that some lenders might experiment with using both types of scores until they are
satisfied as to which score provides the best fit for
different segments of the population.
A related issue concerns where all the data
from the nontraditional credit sources will ultimately be used. According to Horan, if these data are obtained through an exclusive contractual agreement
for the FICO Expansion Score, they will be used
only in connection with the FICO Expansion Score.
However, she noted that some state legislatures are
contemplating whether they should mandate that
these nontraditional data be reported to the major
credit reporting agencies. If that occurs, these data
might also be reflected in the classic FICO score as
well. She pointed out that, in fact, some utility companies are already sending information to the major
credit reporting agencies.13
One attendee questioned the wisdom of using a credit score as part of the employment process.
Horan responded that although the FICO score was
not developed as an employment indicator, she was
well aware that it was being used to make hiring
decisions. However, she noted that this practice did
not represent a significant volume of FICO score
12
www.philadelphiafed.org
Horan gave the example of WE, the Wisconsin Energy
company.
13
Recent Developments in Credit Scoring 13
usage. She also mentioned that a separate version
— the credit-based insurance score — was being
used in decisions about insurance, and such use is
the subject of a study by the Federal Trade Commission. Another speaker on the program added that
using credit scores in employment and insurance
decisions constitutes permissible purposes under the
Fair Credit Reporting Act.
In addressing several questions concerning
the availability of the FICO Expansion Score to consumers, Horan stated that consumers can request a
copy of their Expansion credit report with accompanying explanatory codes and their FICO Expansion
Score from First Advantage, when their alternative
credit information has been scored using this tool.14
IV. VantageScore
Maxine Sweet of Experian introduced the
new VantageScore and also provided some general
information about credit scores and credit scoring
that dovetailed with earlier presentations. Sweet
began her remarks by noting that a de facto scoring
existed even before the first scoring model was developed. During the pre-scoring days, creditors devised
their own “rules-of-thumb” sheets to guide their
lending decisions. They took into account individuals’ financial behavior, such as the number of times
they were 30 or 60 days late in making payments, as
well as the amount of their debt and other factors.
Unfortunately, the lending decision was vulnerable
to the personal bias of the individual evaluating the
credit application. Consequently, the evaluators’ biases toward women, minorities, or other population
subgroups could influence their judgment.
The advent of formal scoring via scoring
models helped to make the lending decision more
objective and curbed the injection of personal bias
Consumers can call 1-866-838-3427 to request a copy of
their Expansion credit report or Expansion score. Although the
FICO Expansion report might be available, the FICO Expansion
Score can be provided only when it has been calculated and
previously used in a lender’s decision process.
14
14 Recent Developments in Credit Scoring
in the process. But Sweet cautioned that although a
credit score is a valuable risk management tool that
supplants a manual scoring sheet, it is only a tool.
It reflects the information in a credit report, which
should be the primary object of concern. While
different scoring models might generate different
credit scores from the same credit report, the fact
remains that what is contained in the credit report is
of utmost importance. Thus, a credit score without
any means of interpretation can be deceiving. She
illustrated this point by referring to her company’s
scoring model known as the Experian Plus Score.
According to Sweet, the original scoring range for
the Experian Plus Score was from 0 to 500, and in
this model, 500 is a bad score — that is, the lower
the score, the better the risk. This ran counter to
the classic FICO score, where a higher score means
a consumer is more creditworthy. The range was
subsequently converted to a scale similar to that
used in the classic FICO score. However, Sweet further noted that many consumers are still under the
misperception that there is only one score.
Sweet observed that although the three major credit reporting agencies (Equifax, Experian, and
TransUnion) partnered with Fair Isaac to develop
their scoring models, these models can use different scoring ranges. This has led to some confusion
among consumers when they receive different credit
scores from different credit reporting agencies. Given
the number of credit scoring models that exist,
Sweet urged that consumers should not be so fixated
on the number per se but instead pay attention to
where their score falls in the model’s range of risk.
This was a convenient segue into Sweet’s
discussion of VantageScore. However, before focusing on VantageScore, she briefly outlined the roles
of the three parties involved in risk scoring. First
are the lenders, who define the lending criteria and
interpret the risk scores. They choose the scoring
model they will use, and some use more than one
model. Moreover, their ranges of approval may vary,
and their product pricing may vary across different
ranges of scores. Second are the risk score modelers.
They develop the underlying algorithm in a model
that produces a credit score. Currently, the largest
and best known is Fair Isaac’s FICO scoring model.
www.philadelphiafed.org
The third party is the credit reporting agencies, such
as Equifax, Experian, and TransUnion, that provide
credit history information on individuals. These credit reports are “scored,” and the credit reporting agencies make both the credit reports and credit scores
available to consumers, along with guidance to assist
consumers in assessing their risk factors.
Sweet indicated that VantageScore is a new
credit score developed by a third-party company in
collaboration with the three major credit reporting
agencies. These entities created VantageScore with
an eye toward having a score that would be highly
predictive but easy for consumers to understand and
lenders to use. The resulting score was designed to
have several desirable features.
First, VantageScore is consistent across the
major credit reporting agencies. Thus, if a consumer’s
information is the same in the credit reports at the
three major credit reporting agencies, his or her VantageScore would be the same at all three agencies.15
Under these circumstances, different scores on the
VantageScore would emerge only if there were differences in the underlying data in the credit reports.
Second, VantageScore’s scale was devised to
be more intuitive. VantageScore’s developers thought
that adopting the equivalent of the familiar grading system of 90-100 = A, 80-89 = B, etc. would
be more meaningful to consumers. Thus, the scale
for VantageScore is 901-990 = A, 801-900 = B,
701-800 = C, 601-700 = D, and 501-600 = F. Corresponding to these ranges are the categories super
prime, prime plus, prime, nonprime, and high risk, respectively.
Sweet also noted that VantageScore provides
more predictive scores for consumers with thin files.
This feature offers more effective risk management to
lenders when working with this segment of consum-
This is because the algorithm that was developed cooperatively
is used by the three credit reporting agencies without modification.
In contrast, the popular FICO scoring model might be applied
differently by the major reporting agencies and result in different
credit scores even if the underlying information is the same.
ers. In addition, she mentioned that VantageScore,
like the FICO score, is based on a consumer’s behavior in the areas of payment history (32 percent),
credit usage (23 percent), balances (15 percent),
types and age of credit (13 percent), recent credit
(10 percent), and available credit (7 percent); it also
informs consumers about both positive and negative
factors that contribute to their score.
Sweet returned to a point she raised earlier, namely, that a number of credit scoring models
in the public domain have different scales. Thus,
consumers should avoid trying to compare different
credit scores from alternative credit scoring systems.
Instead, she advised consumers to focus on the level
of risk a credit score conveys. To underscore this
point, Sweet offered the following example. A consumer obtains his credit scores from three different
sources: a score of 863 from VantageScore, 770 from
Experian, and a 758 score from Equifax. The individual numbers by themselves are not very revealing.
However, once the consumer looks closely at the
ranges associated with each score, the scores’ meaning becomes clear. In this example, the VantageScore
falls in the 801 to 900 range, which is a B, and also in
the top 40 percent. Similarly, since the Experian and
Equifax scores share the same range, they are both
in the 750 to 800 range of risk and are also equal to
the top 40 percent. While Sweet acknowledged that
many consumers should strive to increase their credit
score, they should also be quite cognizant of the
range of risk in which they fall.
In closing, Sweet cautioned attendees that
when it comes to improving a credit score, there is
no silver bullet. Consumers should address the negative risk factors included on their credit report and
allow some time for the changes to be reflected in
a higher score. She observed that consumers don’t
change their credit history overnight; they have to
change their credit usage over time. According to
Sweet, time is the key.16
15
www.philadelphiafed.org
Sweet gave attendees a list of useful resources related to
credit reports and credit scores. These resources can be found
on the Federal Reserve Bank of Philadelphia’s website at: www.
philadelphiafed.org/cca/index.html.
16
Recent Developments in Credit Scoring 15
Discussion
The questions Sweet fielded concerned specific situations that some of the attendees encountered while counseling their clients. One attendee
inquired about the accuracy of credit reports in
light of the “boomeranging” of negative information. Sweet explained that boomeranging typically
occurs when a consumer formally disputes an entry
on a credit report and the credit reporting agency
(CRA) checks with the creditor, who, in turn, fails to
respond (or agrees that the item should be deleted),
prompting the CRA to remove the entry from the
report. However, in the interim, the creditor neglects
to change its records. Thus, when the CRA receives
the creditor’s monthly account updates, the negative
entry reappears on the consumer’s report, resulting in a boomerang. Sweet indicated that Experian
instituted a procedure to deal with this situation.
When Experian deletes an entry from a credit report,
a marker is placed in its system to prevent that entry
from returning. However, she noted that if the creditor provides valid proof — beyond the usual update
— that the negative item should appear on the consumer’s credit report, Experian is required by law to
reinstate it on the consumer’s report and notify the
consumer that it was reinstated.
Another question dealt with whether the
credit report received by a creditor on a potential client contains the same information that appears on
the report obtained by the individual. Sweet pointed
out that there is indeed some information that appears on an individual’s report but that is not included on the report received by a creditor. According to Sweet, a primary example is “soft” inquiries.
She reminded the attendees that these inquiries are
requests for a consumer’s credit report that do not
affect the credit score. She also indicated that a consumer’s account numbers, date of birth, and marital
status do not appear on a credit report received by a
company for employment purposes.
Finally, there was some discussion about the
length of time that a collection item remains on a
credit report and whether a collection agency can
extend the life of the chargeoff on the report. Sweet
and another presenter offered a joint response. They
16 Recent Developments in Credit Scoring
indicated that, in accordance with the Fair Credit
Reporting Act, a collection or chargeoff remains
on a credit report for seven years from the date of
the delinquency that precipitated the chargeoff as
recorded by the original creditor. While some collection agencies would like to restart the clock when
they purchase the charged-off debt, they are prevented by law from doing so. Therefore, the starting
date for the seven-year period remains the original
delinquency date, regardless of any subsequent collection accounts associated with that debt.
V. HealthScore
The final speaker was Jeremy Blackburn
of Canopy Financial, who introduced a new credit
score, the HealthScore. Blackburn set the stage for
the discussion of HealthScore by presenting some
statistics that emphasized the rising cost of health
care and its consequences. He pointed out that
since 2000, the employee share of medical insurance premiums has increased 143 percent and outof-pocket expenses increased 115 percent, while
wages have increased only 15 percent. According
to Blackburn, health-care-related expenses are now
the country’s number one cause of bankruptcies.
Indeed, he reported that every 30 seconds, someone
files for bankruptcy as a result of a serious health
problem.17
Blackburn indicated that various efforts are
underway to restrain the rising cost of health care.
Some of the approaches being taken include the following: investing in technology, such as electronic
medical records, to improve the efficiency of time
management and the administration of patients’
visits to a doctor’s office; using pay for performance
to encourage improvements in the quality of care by
physicians and hospitals with an eye toward reducing costly malpractice suits, and offering high-deductible health plans. He noted that the increased
use of high-deductible plans by consumers is actu-
17
Blackburn further noted that the average out-of-pocket
medical debt for a filer of bankruptcy was $12,000. Both of
these points can be found at: www.nchc.org/facts/cost.shtml.
www.philadelphiafed.org
ally proving to be effective in stemming the tide of
rising health-care expenses.
The deductible in a traditional health insurance plan, according to Blackburn, is inversely related
to the amount of the premium. Thus, to demonstrate
how savings might be realized by switching to a plan
with a higher deductible, he offered the following example. A consumer has a health plan with a $200 deductible, which the consumer is responsible for paying
in full, and a monthly premium of $950, of which the
consumer pays $350 and the employer pays the remaining $600.18 Suppose the consumer elects to have
a higher deductible of $1,200. The monthly premium
is lowered to $295, of which the consumer pays $125
and the employer pays $170. Blackburn points out
that the difference between the monthly premiums
represents real savings enjoyed by both the consumer
and the employer.19 But the consumer now faces a
new, higher financial obligation that accompanies
the higher deductible. Blackburn stressed that the
motivation to move to a high-deductible health plan
is effective only if the employer shares the gain with
the employee. This is often done by placing a portion
of the savings in a tax-advantaged account known as
a health savings account, or HSA. HSAs help employees offset part of the extra expense incurred with
high insurance deductibles. However, if employers are
reluctant to share the gain, employees would have to
bridge the entire gap between the two deductibles on
day one of the high-deductible plan, thus providing a
possible deterrence to adopting the plan.
While there has been considerable growth
in HSAs, Blackburn noted that a sizable gap exists
between those with high-deductible health plans and
no HSAs and those with similar plans but with HSAs.
He sees this gap as an opportunity for financial in-
Blackburn also used this example to underscore the financial
difficulty that employers encounter when offering health benefits
to employees. Thus, the cost to an employer would be $600
multiplied by the number of employees. He asked the attendees to
consider the cost involved if an employer had 10,000 employees,
50,000 employees, or 300,000 employees.
19
In fact, he indicated that the switch to high-deductible health
plans has resulted in employers’ realizing, on average, savings of
40 to 60 percent.
18
www.philadelphiafed.org
stitutions to step in and provide some type of line of
credit that consumers could draw on when they have
unexpected medical problems and face paying medical bills up to their high-deductible amounts. The
availability of such a line of credit might allay consumers’ fears when contemplating a move to high-deductible health plans.
Blackburn then turned his attention to the
discussion of HealthScore. According to Blackburn,
HealthScore can play a vital role in assisting potential lenders in assessing the financial risk associated
with extending credit to consumers who have no
HSAs or insufficient funds in their HSAs to meet
their health plan’s high deductible when they have
medical expenses. He noted that currently credit
reports from the major credit reporting bureaus and
their affiliated credit scores do not include the yearly
financial obligation associated with a consumer’s
health-care plan, even though the plan’s cost (including the deductible) can have a rather significant impact on the consumer’s credit risk profile. However,
Blackburn pointed out that HealthScore gives potential lenders a clearer picture of a consumer’s healthcare plan, thus making a consumer’s creditworthiness
more transparent. He is hopeful that lenders will be
encouraged to make lines of credit available to consumers with health-related financial needs, given the
assistance of HealthScore to assess credit risk and the
existence of regular employer contributions into a
dedicated account as security for a loan — contributions that would come from sharing the employer’s
reduced costs when employees choose a high-deductible plan.
In addition to being a valuable tool to gauge
credit risk associated with health-care lending,
HealthScore, Blackburn noted, affords more transparency when assessing a consumer’s overall risk
picture and thus would be beneficial in underwriting
automobile, credit card, and mortgage loans. This
would be prudent, since health-care expenses are a
cause of 60 percent of all delinquent automobile payments and 45 percent of all delinquent housing payments, whether for rent or mortgage.
Recent Developments in Credit Scoring 17
As a backdrop to further discussing the
virtues of HealthScore, Blackburn’s presentation
included a PowerPoint slide that compared a traditional credit score and the HealthScore for the same
consumer. In his example, the consumer is a full-time
worker with a salary of $45,000, who pays bills on
time and has a $5,000 credit card limit. Based on the
use of retail credit services, the consumer has a traditional credit score of 778. In contrast, a HealthScore
of 662 is attained once the additional information
about the consumer’s high-deductible health plan
is taken into account — namely, a $4,000 deductible and an annual out-of-pocket cost of $10,000.
Even though both the traditional credit score and
the HealthScore are in the same range, Blackburn
stressed that the added data included in HealthScore
helps segment the population to better identify those
most likely to take advantage of credit designed to
cover health-related expenses. He demonstrated
this by pointing out the differences in the location of
profit pools associated with a traditional credit score
and HealthScore. In the case of traditional scores,
the profit pool (or optimal lending window) for those
seeking loans for retail-credit-related services lies in
the range between 711 and 791, while the optimal
profit pool for HealthScore is between 590 and 711.
In view of the billions of dollars of bad
debt incurred by hospitals and medical providers,
Blackburn suggested that a consumer armed with a
HealthScore might be in a better position to negotiate discounts for health-care services from a healthcare provider. Furthermore, he emphasized that the
added information in HealthScore can assist lenders
in pricing their health-care-specific line of credit by
better determining the parameters of the loan terms.
Blackburn concluded his presentation with
two overriding points he wanted to make to the symposium’s attendees. First, he wanted to impress upon
them that, notwithstanding its virtues, HealthScore
should not be regarded as a stand-alone score. Instead, Blackburn thinks that HealthScore should be
an essential input in the VantageScore, FICO Expansion Score, or the traditional FICO score. In particu-
18 Recent Developments in Credit Scoring
lar, he believes that the responsible management of a
high-deductible health-care plan with its accompanying HSA should be included as a positive entry on
a consumer’s credit report, particularly consumers
with low incomes or thin files. Second, Blackburn
wanted to emphasize the fact that, ultimately, a
credit score is only a number. However, given the
impact a credit score can have on a consumer, he
urged that it be based on the major factors affecting a consumer’s creditworthiness. Yet, information
pertaining to health insurance, which is a key influence on a consumer’s credit profile, is not currently
reflected in any of the credit scores of the major
credit reporting agencies. According to Blackburn,
HealthScore provides lenders with the necessary
information about the influence of health insurance
when assessing lending risk.
Discussion
Blackburn entertained several questions, all
of which dealt with some aspect of the data used in
HealthScore or how it is derived. Without divulging any proprietary information, he offered general
responses that addressed the questions. Blackburn
pointed out that the data used to construct HealthScore are gleaned from the information a consumer
provides on an application for health insurance coupled with the specifics of the health plan’s design.
Of particular importance are the plan’s deductible,
premium, and maximum out-of-pocket costs. These
data are augmented with the consumer’s income,
employment status, and length of employment. Employment status serves to distinguish between fulland part-time employment, since employers do not
contribute to HSAs for part-time employees. Moreover, the length of employment gauges the likely
balance in an HSA. Without getting into specifics
of the derivation of HealthScore — since it is proprietary — Blackburn indicated that HealthScore
uses ratios of certain items to income, and these ratios help to scale the importance of these factors in
terms of their affordability to the consumer. Blackburn also stressed that HealthScore is not based on
any medical data.
www.philadelphiafed.org
VI. Concluding Note
Consumers should be aware that a credit
score is a very powerful number. Consequently, it
is incumbent on them to know as much as possible
about credit scores and credit scoring. The symposium introduced the attendees to three new creditscoring products that will no doubt play a role in
the financial arena. Those in attendance heard that
VantageScore offers greater consistency in scoring
methods among the three major credit reporting
agencies; the FICO Expansion Score uses nontraditional data to increase the number of individuals
www.philadelphiafed.org
receiving scores; and the HealthScore incorporates
health-plan and other information in order to provide the health-care industry and lenders with an
instrument to better gauge an individual’s overall
credit risk.
All told, the symposium was very informative. The attendees left with a newfound understanding of credit reports and credit scores, as well
as insights into new credit score instruments that
will possibly affect their lives and the lives of the
constituents they serve.
Recent Developments in Credit Scoring 19
Exhibit 1:
Symposium Agenda
9:00 a.m.
Continental Breakfast
9:30 a.m.
Welcome.
Dede Myers, Federal Reserve Bank of Philadelphia
Overview of Credit Scoring
Patricia Hasson, Consumer Credit Counseling Service of Delaware Valley
9:35 a.m.
10:15 a.m.
FICO Expansion Score
Janice Horan, Fair Isaac Corporation
10:55 a.m.
Break.
VantageScore.
Maxine Sweet, Experian
11:10 a.m.
11:50 a.m.
HealthScore.
Jeremy Blackburn, Canopy Financial
12:30 p.m.
Lunch.
www.philadelphiafed.org
Recent Developments in Credit Scoring 21
Exhibit 2:
Institutions Represented at the Symposium
AHOME, Inc.
American Credit Alliance, Inc.
Asian Bank
Beneficial Savings Bank
Bucks County Housing Group
Canopy Financial
CCCS of Delaware Valley
Citicorp Trust Bank, FSB
Citizens Mortgage Corp
Community Action Agency of Delaware County, Inc.
Community Action Commission
Community First Fund
Consumer Credit Education & Protection Cooperative
Cooperative Business Assistance Corporation
Delaware County Office of Housing & Community Development
Delaware County OHCD
Delaware Money Management Program
Excalibur Financial, Inc
Experian
Fair Isaac Corporation
Federal Reserve Bank of Philadelphia
Fern Rock-Ogontz-Belfield CDC
Fidelity Savings of Bucks County
First Keystone Bank
Franklin Mint Federal Credit Union
Fulton Bank
Fulton Financial Corporation
Genesis Housing Corporation
Greater Philadelphia Urban Affairs Coalition
Housing Association
Housing Partnership of Chester County
ING Direct
JEVS Human Resources
www.philadelphiafed.org
Recent Developments in Credit Scoring 23
Institutions Represented at the Symposium
JP Morgan Chase
KPMG Banking Insider
Liberty Resources
LISC
Malvern Federal Savings
Maryland Cooperative Extension
Media Fellowship House
MidPenn Legal Services
Mortgage and Credit Center
National Penn Bank
NCALL Research, Inc.
Neighborhood House, Inc.
Neighborhood Transformation Initiative/EZ
NeighborWorks America
New Kensington CDC
Newfield National Bank
Norris Square Civic Association
OHCD - City of Philadelphia
PA Department of Banking
PA Department of Revenue
PathWaysPA
PBCIP, Inc.
Penn State Cooperative Extension
Pennsylvania Housing Finance Agency
Pennsylvania Insurance Department
Philadelphia Federal Credit Union
Project H.O.M.E.
Select Bank
Southwest CDC
Sovereign Bank
St. Edmond’s Federal Savings Bank
Sterling Financial Corporation
Susquehanna Patriot Bank
24 Recent Developments in Credit Scoring
www.philadelphiafed.org
Institutions Represented at the Symposium
Temple University School of Podiatric Medicine
The Housing Association Information Program
The Partnership CDC
The Philadelphia Trust Company
Third Federal Bank
Transitions
www.philadelphiafed.org
Recent Developments in Credit Scoring 25
Ten Independence Mall
Philadelphia, PA 19106-1574
215-574-6393
215-574-7101 (fax)
www.philadelphiafed.org
Marvin M. Smith, Ph.D.
Community Development Research Advisor
Community Affairs Department
26 Recent Developments in Credit Scoring
www.philadelphiafed.org